Logistic Regression with Count Vectorizer

[ian-flores]

First Approach:

1) Count number of words 2) Train Logistic Regression on this only

Second Approach:

1) Remove Stopwords 2) Count number of words 3) Train LR

Third Approach:

1) Remove Stopwords 2) Count number of words 3) Weight number of words per document with tf-IDF 4) Train LR

[ian-flores]

First Approach:

• Training Accuracy: 0.9187821174111318
• Testing Accuracy: 0.9178692927484333

Second Approach:

• Training Accuracy: 0.9201511747862053
• Testing Accuracy: 0.918531781557744

Third Approach:

• Training Accuracy: 0.9592753637161231
• Testing Accuracy: 0.9544136078782453